Researches Of Gaze Tracking Calibration Technique Based On Eye Position

With the rapid development of computer science and technology, computer technology has been applied into the fields of multi-media and man-machine interaction, the rapid growth of computer users are concerned intelligent human-computer interaction. In the early period, gaze tracking technology was mainly used in psychological research and medical diagnosis, and at present eye tracking technology has already been used to more areas with the rapid development of computer technology and multimedia networks.For example, in human-computer interaction, gaze tracking is used to input character and draw; in medicine, in addition to medical diagnosis and psychological research, patients with active inflexible also can control appliances around based on gaze tracking technology; In business, gaze tracking technology is also widely used in the assessment and web advertising testing, to achieve the optimal advertising and web designing; In the military field, gaze tracking technology can be precision-guided. Gaze tracking system plays an intelligent potential in human-computer interaction, and it is a hot research.Although the application of gaze tracking technology has tempting prospects and it has achieved great successes in the last few years, the gaze tracking technology cannot be widely used in the actual production due to the immaturity of various techniques, characters of eye movement itself and the high cost. Particularly, most gaze tracking technology needs the process of calibration, which is an inconvenience for users. In addition, some problems have to be solved, which include improving the accuracy of the gaze tracking system, improving the degree of freedom, reducing the complexity of calibration, reducing the complexity of algorithm and reducing the cost of the system. The main contributions in our paper are shown as following:(1) The estimation errors of the traditional binocular camera based gaze tracking algorithm are analyzed in3D space. The analysis results show that the errors have regional characteristic. It provides the theoretical basis for the system calibration and improvement of the accuracy. (2) Based on the characteristics of the human eye, the deviation between the optical axis and the visual axis are analyzed. The analysis results show that the deviation between the optical axis and visual axis has certain regularity. It provides the theoretical basis for the system calibration.(3) Based on the principle of the error generation, the relationship between the deviation that between the optical axis and the visual axis and the eyes’ position is analyzed. The algorithm of gaze tracking calibration technique with eye position is brought. It improves the accuracy and the degree of freedom further.The proposed method makes the computer to do the calibration process, and it promotes the application of the gaze tracking techniques to be more extensive. Users can control the system naturally.